Thyroid hormone replacement for nephrotic syndrome patients with euthyroid sick syndrome: a meta-analysis

Abstract

Objective: To assess the efficacy of thyroid hormone replacement therapy for nephrotic syndrome (NS) patients associated with euthyroid sick syndrome (ESS). Materials and methods: The Cochrane library, ISI, Ovid, PubMed, Chinese Biomedicine Database were searched, and reference list of relevant articles were selected. Randomized controlled trials (RCTs) or quasi-RCTs with thyroid hormone replacement on NS patients associated with ESS were included in this analysis. Results: Six trials (329 participants) were included. Meta-analysis showed that thyroid hormone replacement therapy can significantly increase the completely remission rate [OR = 3.04, 95% confidence interval (CI): 3.04–1.88, p < 0.00001] and total response rate (OR = 4.63, 95% CI: 2.46–8.71, p < 0.00001) of NS patients associated with ESS. No side effect was observed during the follow-up period. There was no obvious publication bias in the mete-analysis studies. Conclusions: Thyroid hormone replacement therapy significantly increases the remission of ESS in patients with NS.

• Thyroid hormone replacement
• nephrotic syndrome
• euthyroid sick syndrome
• meta-analysis

Introduction

Euthyroid sick syndrome (ESS), which is also named nonthyroidal illness syndrome (NTIS) or low triiodothyronine (T3) syndrome, is characterized by decreased serum T3 and thyroxin (T4) levels in critical or noncritical illness, accompanying with increased reverse T3 (rT3) and no significant increase in thyroid-stimulating hormone (TSH).1 It is usually not caused by the thyroid or pituitary gland diseases, but probably a response to the underlying illness,2,3 including severe infections,4,5 trauma,6 major surgery,7,8 myocardial infarction,9 inflammatory conditions,10,11 starvation12 and kidney diseases.13 In most cases, patients with ESS have no peculiar or obvious symptoms, thus ESS might be neglected by doctors. Studies showed that ESS is not uncommon in hospitalized patients, especially patients in the intensive care units,14,15 and it may be a sign of a poor prognosis and may contribute to mortality in critical patients.6

It is still a controversy whether thyroid hormone replacement therapy should be used to these patients. Pappa et al. thought that there is no clear evidence of benefit from thyroid hormone replacement and well-designed studies that confirming its efficacy, thyroxine supplementation should not be recommended for the treatment of NTIS.16 While De Groot et al. hold that ESS is a manifestation of hypothalamic–pituitary dysfunction and should be treated with appropriate replacement therapies.17

In this study, we hope to explore the efficacy and safety of thyroid hormone replacement therapy for nephrotic syndrome (NS) patients associated with ESS. We collected related articles extensively, strictly selected eligible studies, evaluated the efficacy of thyroid replacement therapy for NS patients associated with ESS by the method of meta-analysis, so as to provide guidelines for clinics.

Materials and methods

Literature search

By methods of MeSH, keywords and manual retrieval, we searched the Cochrane library, ISI, Ovid, PubMed and Chinese Biomedicine database. The search terms includes NS, thyroxine/thyroxin/levothyroxine/thyroid hormone, ESS, nonthyroid/NTIS and ESS/NTIS. There were no language and publish date restrictions. Related research references were retrieved at the same time.

Inclusion criteria

For inclusion, trials need to satisfy the following criteria: (a) patients definitely diagnosed with NS and ESS, (b) study design should be randomized controlled trials (RCTs) or quasi-RCTs and (c) complete data and clear definition of completely remission (CR) and partially remission (PR).

Exclusion criteria

Trials were excluded when meet one of the following items: (a) no controls, (b) no complete data, (c) data or papers published repeatedly, (d) subjects have a history of thyroid or pituitary diseases and (e) use of drugs that affect the function of pituitary or thyroid.

Study quality

The quality of included studies was assessed by two researchers independently according to the Jadad standard.18 The Jadad scale includes the following items: (a) with or without randomized methods, (b) appropriateness of randomization methods, (c) with or without double-blind methods, (d) appropriateness of blinding methods and (e) with or without records of incidence of withdraw and dropouts and the reason. The Jadad score ranges from 0 to 5 points, with 0–2 belonging to low quality, 3–5 belonging to high quality.

Data extraction and outcome measures

The process of data extraction of included studies was conducted by two researchers independently. If they disagree with the extracted data, it will be solved through discussion or contacting the author for further information. Extracted information contains the first author, year of publication, the side effects, definition of completely and PR, the follow-up time, the number of participants and the number of completely and partially remitted patients in experimental and control groups. CR: urine protein <0.3 g/24 h with normal kidney function, no relevant symptoms. PR: urine protein 0.3–1.0 g/24 h with normal kidney function, no relevant symptoms. Invalid: urine protein >1.0 g/24 h with or without worsening kidney function, no remission of symptoms.

Statistical analysis

Statistical analysis of included studies was performed with RevMan5.0 software provided by the Cochrane collaboration (The Informatics & Knowledge Management Department, CochraneTech). I² test and Q statistic test were used to examine the heterogeneity, with I² > 50% or p < 0.10 considered to be significant.19 If I² > 50% or p < 0.10, the random effects model was used for meta-analyses, otherwise, the fixed effects model was selected. The difference of effects between experimental and control group was expressed as odds ratio (OR), as well as its 95% confidence interval (95% CI). Publication bias was examined by funnel plot method along with meta-analysis.

Results

Studies included in the meta-analysis

Thirteen relevant articles were identified according to the inclusion and exclusion criteria and six articles (329 patients)20–25 were finally included in this study. The flow diagram of selection of articles is shown in Figure 1, and the characteristics of each study are listed in Table 1. All the patients of the six studies were treated with glucocorticoid as a major, with 1 mg/kg·d of prednisone as a starting dose.

Figure 1. Selection of the studies included in the meta-analysis.

Table 1. Characteristics of included studies.

Studies	Design	Jadad scores	Follow-up (months)	Differences in therapies	Number of participants	Results
Duan et al., 200821	Randomized, double-blind, parallelled	4	12	Experimental group: levothyroxine 50 µg/(kg d); Controls: placebo	40 in experimental group, 38 in controls	Experimental group 34 CR, 3 PR, 3 invalid; Controls 26 CR, 4 PR, 8 invalid
Hu et al., 200522	Randomized, double-blind, parallelled	4	12	Experimental group: levothyroxine 50 µg/(kg d); Controls: placebo	26 in experimental group, 24 in controls	Experimental group 22 CR, 2 PR, 2 invalid; Controls 14 CR, 3 PR, 7 invalid
Li et al., 200523	Randomized, double-blind, parallelled	4	3	Experimental group: levothyroxine 50 µg/(kg d); Controls: placebo	25 in experimental group, 23 in controls	Experimental group 16 CR, 6 PR, 3 invalid; Controls 11 CR, 3 PR, 9 invalid
Nie et al., 200824	Randomized, double-blind, parallelled	3	6	Experimental group: levothyroxine 25–50 µg/(kg d); Controls: placebo	36 in experimental group, 30 in controls	Experimental group 30 CR, 3 PR, 3 invalid; Controls 15 CR, 5 PR, 10 invalid
Zhang et al., 200425	Randomized, double-blind, parallelled	4	6	Experimental group: thyroxine 20–40 mg/(kg d); Controls: placebo	26 in experimental group, 24 in controls	Experimental group 16 CR, 8 PR, 2 invalid; Controls 11 CR, 7 PR, 6 invalid
Xie, 200920	Randomized, double-blind, parallelled	3	12	Experimental group: levothyroxine 20 µg/(kg d); Controls: placebo	20 in experimental group, 17 in controls	Experimental group 16 CR, 2 PR, 2 invalid; Controls 7 CR, 3 PR, 7 invalid

Notes: CR: completely remission and PR: partially remission.

Effects of thyroid hormone replacement therapy on CR rate

Six studies involved 329 subjects, with 173 in the experimental group and 156 in the control group. After treatment, 134 of 173 subjects in the experimental group and 84 of 156 subjects in control group recovered completely. Meta-analysis shows that there is no significant heterogeneity between six studies (p = 0.71, I² = 0%), thus, the fixed effects model is selected. The difference of effect between experimental group and control group is statistically significant (OR = 3.04, 95% CI: 1.88–4.94, p < 0.00001), as shown in Figure 2A.

Figure 2. (A) The meta-analysis of completely remission rate of six studies (fixed effects model). (B) The funnel plots of six studies in analysis of completely remission rate (fixed effects model).

Effects of thyroid hormone replacement therapy on total response rate

One-hundred fifty-eight of 173 subjects in the experimental group and 109 of 156 subjects in control group responded to the therapy. Meta-analysis shows that there is no significant heterogeneity between six studies (p = 0.99, I² = 0%). Thus the fixed effects model is selected. The difference of effect between experimental groups and control groups is statistically significant (OR = 4.63, 95% CI: 2.46–8.71, p < 0.00001), as shown in Figure 3A.

Figure 3. (A) The meta-analysis of rate of total response rate of six studies (fixed effects model). (B) The funnel plots of six studies in analysis of total response rate (fixed effects model).

Side effects

All six studies have no reports on side effects of experimental and control groups; therefore, we cannot evaluate the potential side effects.

Publication bias

The funnel plots are shown in Figures 2B and 3B. The pattern of funnel plots for CR rate and total response rate are roughly symmetrical for indicating the publication bias.

Discussion

Under physiological conditions, the serum levels of thyroid hormone is closely related to the homeostatic regulation of the hypothalamus–pituitary–thyroid axis in which the hypothalamic thyrotropin-releasing hormone (TRH) stimulates the secretion of TSH by the anterior pituitary, thus resulting in the secretion of thyroid hormone by the thyroid gland.26 The homeostasis regulation of serum thyroid hormone can be disordered in illness or surgical stress conditions, partially resulting in ESS, which with an important influence on the prognosis.15

The mechanism of ESS in NS patients remains unclear. The studies of Fonseca V et al. showed that among the NS patients with urine T4 positive have a lower serum T3 and T4,27,28 suggesting that urinary thyroid hormone loss may play a pivotal role in lower levels of serum T3 and T4. Massive proteinuria may result in loss of thyroid-binding globulin (TBG)29 and furosemide can compete for T4-binding sites on TBG, both contributing to lower total T4 and T3 levels.30 While oxidative stress, psychological pressure and the long-term use of glucocorticoid can inhibit the expression of TSH in pituitary, balancing out the elevated TSH caused by the low levels of serum thyroid hormones, thus making the levels of serum TSH be normal or slightly low.31–33 Yu et al.34 showed that the syndrome was due, in part, to the decreased activity of type I 5′-deiodinase (5′-DI), the hepatic enzyme that converts T4 to T3. By primary culturing of rat hepatocytes, they demonstrated that the expression of 5′-DI was suppressed at the transcriptional level in the condition of ESS, with interleukin-1 and interleukin-6 inhibit the induction of 5′-DI RNA and enzyme activity.

It is still unclear whether ESS represents a physiological protective mechanism, or a damaging maladaptive response, and whether early thyroid hormone administration might improve the prognosis in ESS.35 Actually, the low level of serum thyroid hormone can inhibit the expression of glucocorticoid receptor (GR), with manifestations of steroid resistance.36 The elevated serum thyroid hormone levels after thyroid hormone replacement therapy increases the expression of GR, enhancing the binding of the GR and glucocorticoid, ultimately strengthening the efficacy of glucocorticoid for NS. Thyroid hormone plays a pivotal role in water and sodium homeostasis, renal development and renal hemodynamics.37–39 The renal blood flow was increased after thyroid hormone replacement therapy, promoting the restoration of glomerular basement membrane and the remission of NS.

The meta-analysis shows that thyroid hormone replacement therapy significantly improved the remission of the NS patients with ESS not only in CR rate (OR = 3.04, p < 0.00001) but also in total response rate (OR = 4.63, p < 0.00001) with no records of adverse effects in the process of treatment. No publication bias was detected in this meta-analysis.

As the sample size is relatively small, differences may exist between the results of meta-analysis and that of the actual clinical situations. Therefore, the conclusion should be treated cautiously. For more credible conclusion, studies of multicenter and larger sample sizes are needed.

Declaration of interest

No conflict of interests is declared.

This research was supported by the National Natural Science Foundation of China (no. 81360122/H0518).

This paper is not under consideration elsewhere. None of the paper’s contents has been previously published. All the authors have read and approved the manuscript.